基于群智能算法优化LSTM的催化裂化预测研究  

Prediction of Catalytic Cracking Based on Swarm Intelligence Algorithm Optimization of LSTM

作　　者：洪娟 田文德[1] Hong Juan;Tian Wende(College of Chemical Engineering,Qingdao University of Science&Technology,Qingdao 266042,China)

机构地区：[1]青岛科技大学化工学院,山东青岛266042

出　　处：《山东化工》2023年第18期92-96,共5页Shandong Chemical Industry

基　　金：国家自然科学基金(22178189);山东省自然科学基金(No.ZR2021MB113)。

摘　　要：针对长短期记忆神经网络(LSTM)预测模型中许多网络参数过分依赖于经验设置,人工参数设置导致模型的精度低、泛化能力弱等问题,采用搜索范围广、收敛速度快的粒子群算法(PSO)和布谷鸟算法(CS)对LSTM的一些超参数进行优化,构建PSO-LSTM模型和CS-LSTM模型,寻找到LSTM的最优参数集,从而更好地提高了模型预测精度。将优化后的模型应用于催化裂化吸收稳定系统主要控制变量解吸塔再沸器返塔温度预测中,验证了模型的有效性。In order to solve problems that many network parameters of long short-term memory network(LSTM)prediction model rely too much on empirical setting,and artificial parameter setting leads to low accuracy and weak generalization ability of the model,particle swarm optimization(PSO)and Cuckoo algorithm(CS)with wide search range and fast convergence speed are used to optimize some super parameters of LSTM.The PSO-LSTM model and CS-LSTM model are constructed to find the optimal parameter set of LSTM,so as to better improve the prediction accuracy of the model.The optimized model is applied to the main control variable of the absorption stabilization system in catalytic cracking to predict the return temperature of the desorption tower reboiler,and the effectiveness of the model is verified.

关 键 词：长短期记忆神经网络 粒子群算法 布谷鸟算法 催化裂化过程 预测 

分 类 号：TE65[石油与天然气工程—油气加工工程] TH3[机械工程—机械制造及自动化]